Domain 3Security Architecture & EngineeringCORE

Cloud Computing & Virtualization Security

NIST cloud deployment and service models (IaaS/PaaS/SaaS), shared responsibility matrix, virtualization security, hypervisor vulnerabilities, and CASB.

Domain: Domain 3: Security Architecture and Engineering•Updated: 2026-08-26
#Cloud Security#IaaS#PaaS#SaaS#Shared Responsibility#Hypervisors#CASB#Containers

CISSP Domain 3: Cloud Computing & Virtualization Security — Mindmap Review Study Guide

═══════════════════════════════════

PART 1: MERMAID MINDMAP

═══════════════════════════════════

mindmap
  root((Cloud Computing))
    Intro and Context
      S3 Buckets
      CCSP Certification
    Characteristics
      On-Demand Self Service
      Shadow IT
      Broad Network Access
      Resource Pooling
        Compute Resource
        Storage Resource
        Network Resource
      Rapid Elasticity
      Measured Service
    Service Models
      IaaS
      PaaS
      SaaS
      Containers as a Service
      Serverless FaaS
      SOC 2 Reports
    Deployment Models
      Public Cloud
      Private Cloud
      Community Cloud
      Hybrid Cloud
    Virtualized Compute
      Virtual Machine Instance
      Hypervisor VM Monitor
      Type 1 Hardware Hypervisor
      Type 2 Software Hypervisor
      Containers
      Containerization Engine
      Insecure Authorization Vulnerability
      Monolithic Application
      Microservices
      Functions as a Service
    Identity in Cloud
      Identity as Perimeter
      Zero-Trust Model
      Local Identities Active Directory
      Cloud Identities Okta
      Cloud Identity Type
      Linked Identities
      Synced Identities
      Federated Identities
    Roles and Responsibilities
      Accountable
      Responsible
      Cloud Consumer
      Owner Data Controller
      Cloud Provider Processor
      Cloud Broker
        Aggregation Arbitrage Intermediation
      Cloud Auditor
    Cloud Protocols
      SPML
      SAML
      OpenID
      OAuth
    Migration and Contracts
      Data Centric View
      Data Lifecycle Stages
      SLA
    Cloud Forensics
      VM Snapshot
      Virtual Disk Volumes
      VM Image
    Data Destruction
      Defensible Destruction
      Crypto Shredding AES

═══════════════════════════════════

PART 2: TECHNICAL DEFINITIONS & CONCEPTS

═══════════════════════════════════

Intro and Context

Cloud Computing — The practice of moving infrastructure, applications, and data to the cloud, starting with Amazon’s S3 in March 2006. It is defined by a specific set of characteristics rather than a single simple definition.

Amazon’s Simple Storage Service (S3 Buckets) — The first cloud service that appeared in March 2006, marking the beginning of massive cloud adoption.

CCSP (Certified Cloud Security Professional) — A separate certification created by ISC2 focused specifically on cloud security, which is why the CISSP exam contains relatively few cloud-specific questions despite the technology’s importance.

Characteristics

Cloud Computing Definition — Defined by a set of characteristics; if a service complies with these characteristics, it is probably cloud.

On-Demand Self Service — Users can request services and the cloud service provider automatically provisions them, usually within milliseconds or seconds, allowing rapid innovation and the ability to fail fast.

Shadow IT — The dark side of On-Demand Self Service, where users provision powerful services without centralized IT approval or oversight.

Broad Network Access — Access to cloud resources is available from multiple device types and multiple locations, primarily through a web browser across a network.

Resource Pooling — A defining characteristic where the three primary technology resources are pooled, shared, and accessed via a layer of virtualization rather than direct physical access.

Compute Resource — The ability to fetch instructions, decode them, execute them, and store the results (essentially the ability to execute code).

Storage Resource — The ability to store bits (data and code) somewhere.

Network Resource — The ability to move bits around everywhere.

Rapid Elasticity and Scalability — The ability to quickly provision and deprovision resources (like vast amounts of bandwidth or terabytes of storage), allowing rapid increase and decrease of cloud service usage.

Measured Service — The cloud service provider closely monitors usage so the consumer only pays for what is used; resource usage is monitored and reported to the consumer, providing visibility and transparency of rates and costs.

Service Models

Infrastructure as a Service (IaaS) — An environment where customers deploy virtualized infrastructure (servers, appliances, storage, networking), allowing them to recreate an entire physical data center as virtualized components (e.g., virtual firewalls, routers, servers).

Platform as a Service (PaaS) — Provides the services and functionality for customers to develop and deploy custom applications without having to worry about underlying complexity like servers, network, and storage.

Software as a Service (SaaS) — A model where a customer rents access to an application hosted in the cloud.

Containers as a Service (CaaS) — A service model that fits between Infrastructure as a Service and Platform as a Service.

Serverless / Functions as a Service (FaaS) — A service model that fits between Platform as a Service and Software as a Service.

SOC 2 Reports — Service level reports commonly used to provide assurance that a cloud service provider has controls in place operating effectively to meet defined requirements.

⚠️ EXAM TIP: The customer ultimately remains accountable for the protection of any data and services outsourced to the cloud. The onus is on the customer to ensure responsibilities are clearly defined in contracts and SLAs, and to verify the CSP’s controls (often via SOC 2 reports).

Deployment Models

Public Cloud — Cloud services available to the general public, where a cloud service provider owns and operates the infrastructure open for public use.

Private Cloud — Cloud infrastructure provisioned for exclusive use by a single customer; can be owned/operated by the customer or CSP, exist on or off premise, and be physically or logically separated from other customers.

Community Cloud — Cloud infrastructure accessible only by a small community of organizations that have similar shared concerns (e.g., similar security and regulatory requirements).

Hybrid Cloud — A combination of public, private, and community cloud models (e.g., using an on-premise private cloud for sensitive data and the public cloud for less sensitive workloads).

Virtualized Compute

Virtual Machine (VM Instance / Guest) — An operating system and applications running on top of a layer of abstraction (hypervisor) instead of directly on physical hardware.

⚠️ EXAM TIP: Virtual machines might also be referred to as Instances or Guests on the exam.

Hypervisor (VM Monitor) — A piece of software providing a layer of abstraction between hardware and the OS; acts as a “giant liar” by simulating underlying hardware to each OS so multiple OSs can run simultaneously without conflicting over total hardware control.

Type 1 Hypervisor — Hardware hypervisors; more secure and efficient than Type 2.

Type 2 Hypervisor — Software hypervisors.

⚠️ EXAM TIP: Make sure you know the difference between Type 1 (Hardware) and Type 2 (Software) hypervisors. Type 1 is more secure and efficient.

Containers — Highly portable code execution environments that run within an operating system, sharing and leveraging resources of that OS; code is packaged inside to run anywhere (Mac, Linux, Azure, Google Cloud) without repacking.

Containerization Engine — The layer of abstraction located between the operating system and the containers (unlike a hypervisor, which sits between hardware and the OS).

Insecure Authorization Vulnerability — A vulnerability referring to doing a poor job of the authorization step, potentially allowing an attacker to bypass authorization or grant themselves unauthorized access.

  • To prevent this, authorization should be performed by the back-end server, not the mobile device, and the server must verify requests are permissible based on what the user is authorized to access.

Monolithic Application — A historical application development approach where all functions are packaged together into a large executable, making it difficult to scale performance as demands grow.

Microservices — An approach where an application is broken down into groupings of functionality, making it easier to scale specific parts of the application as needed.

Functions as a Service (Serverless) — An approach where an application is broken down into individual simple functions that talk to each other via API calls, stored in the cloud, and called as much or as little as desired (paying nothing if never called).

Compute Type Layer of Abstraction Location Execution Environment
Virtual Machine Between hardware and OS Full OS and apps running on simulated hardware
Container Between OS and Container Code execution environment sharing/leveraging host OS resources
Serverless (FaaS) N/A (Developer doesn’t manage infrastructure) Individual simple functions talking via API calls

Identity in Cloud

Identity as the New Perimeter — The concept that moving to the cloud destroys the well-defined perimeter and trusted internal network, making identity the primary security boundary.

Zero-Trust Model — A security approach assuming all traffic is a potential threat with no trusted internal network, requiring robust identification, authentication, and authorization controls.

Local Identities — User identities stored on-premise in the organization’s own data center (e.g., Active Directory).

Cloud Identities — User identities stored using a cloud service (e.g., Okta).

Cloud Identity (Type) — An identity created and managed solely in the cloud.

Linked Identities — Two separate identities (one cloud, one local) with an indication of linkage, but changes to one are NOT automatically synchronized to the other.

Synced Identities — Two separate identities (one cloud, one local) that are synchronized; a change to one is automatically reflected in the other.

Federated Identities — A single identity that allows a user to gain access to both local and cloud-based services via Federated Access.

Identity Type Location Synchronization
Cloud Identity Solely in the cloud N/A
Linked Identities One cloud, one local NOT automatically synchronized
Synced Identities One cloud, one local Automatically synchronized
Federated Identities Single identity for both Accesses both local and cloud via Federated Access

Roles and Responsibilities

Accountable — The individual with ultimate ownership, answerability, blameworthiness, and liability for an asset; assigned to ONLY ONE person per asset because they are “the throat that gets choked if something goes wrong.” Accountability CANNOT be delegated.

Responsible — The doer(s) in charge of executing the requirements defined by the accountable person; can be assigned to multiple people and CAN be delegated.

Cloud Consumer — The customer (person or organization) using cloud services.

Owner / Data Controller — Individuals within the cloud consumer who own the data; they are accountable for the protection of any data stored/processed in the cloud and cannot delegate this accountability.

Cloud Provider / Processor — The cloud service provider; responsible for protecting consumer data based on the owner’s requirements, and accountable for running their own infrastructure and protecting their own data.

Cloud Broker — Middlemen organizations existing between the consumer and CSP to perform Aggregation, Arbitrage and Intermediation (packaging services or adding functionality).

Cloud Auditor — Individuals who audit to see if controls are properly designed and operating effectively.

⚠️ EXAM TIP: Accountability CANNOT be delegated, but responsibility CAN be delegated. The owner/data controller remains accountable for data protection even when outsourcing to a cloud provider.

Cloud Protocols

SPML (Service Provisioning Markup Language) — An XML-based framework for exchanging provisioning information (setup, change, revocation of access) between cooperating organizations to standardize and simplify provisioning access across multiple systems.

SAML (Security Assertion Markup Language) — A protocol that provides both Authentication and Authorization in Federated Access.

OpenID — A protocol that provides only Authentication.

OAuth — A protocol that provides only Authorization capabilities.

Protocol Primary Function
SPML Exchanging provisioning information (setup, change, revocation)
SAML Authentication AND Authorization in Federated Access
OpenID Authentication ONLY
OAuth Authorization ONLY

Migration and Contracts

Data Centric View — A migration approach focusing on data classification and the security controls needed for each stage of the data lifecycle.

Data Lifecycle Stages — The exact stages of data that must be secured during migration: creation, storage, use, sharing, archiving and destruction.

SLA (Service Level Agreement) — Documented commitments by the Service Provider to a Consumer covering confidentiality, integrity, availability, responsiveness, etc.; serves as an addendum to the overall Contract.

Cloud Forensics

VM Snapshot — A copy of a VM preserving the state and data at a specific point in time, including volatile data (CPU cache/register, RAM); the most useful source of evidence.

Virtual Disk / VM Volumes — Virtual hard drives containing lots of data but none of the volatile data; less useful for forensics than a snapshot.

VM Image — A pre-built virtual machine useful for identifying vulnerabilities/misconfigurations, but does not contain any data of the running system; the least useful for forensics.

Evidence Type Volatile Data Included? Forensic Usefulness
VM Snapshot Yes (CPU cache, RAM) Most useful
Virtual Disk / Volumes No Less useful
VM Image No (pre-built, no running data) Least useful

Data Destruction

Defensible Data Destruction — A requirement under many laws/regulations for a data owner to ensure sensitive/personal data is defensibly and demonstrably destroyed.

Crypto Shredding / Crypto Erase — A method of defensible destruction where sensitive data is encrypted with an excellent algorithm like AES, and then every single copy of the encryption key is physically destroyed, making the data unrecoverable and defensibly destroyed.